I agree with @NateSilver538 that an electoral-college strategy build around MI, WI, PA in 2020 is looking very viable after Tues. fivethirtyeight.com/features/the-2… But I'd add a word of caution. (Thread)
2) In a nutshell: It’s still likely to be pretty damn close in those states. Nothing suggesting a blue wall—that’s very much the wrong metaphor. I’d say more like a very slight Dem skew. (The second map Nate posits, not the first.) 
3) Here’s why: As I mentioned in my previous Michigan thread, Dems did very well on Tuesday with two groups: They ran up the score with affluent women, and they limited their losses among working-class whites.
4) Overall, Dems improved their performance among college-educated white women by 20 points on Tuesday (from a 6-point win in 2016 to a 26-point win on Tues).
5) They improved their performance among non-college grad whites by 19 points on Tues (losing by 31 in 2016, but only by 12 on Tuesday).
6) But there are a few problems here for Dems when you look to 2020.
7) State Republicans are pretty unpopular among white working-class voters in Michigan. They’ve been very anti-union. They taxed pensions. They used a procedural trick to try to undermine a min-wage increase. They’ve allowed the roads to deteriorate badly. nytimes.com/2018/09/16/us/…
8) Trump, so far as I could tell in my reporting, is still semi-popular among a lot of white working class folks, who really like his tariffs, his China-bashing, his renegotiation of NAFTA.
9) Which is to say, it’s possible that, in Michigan, voters weren’t rejecting Trump on Tues; they were rejecting state Republicans. (Same goes for Wisconsin, where Scott Walker was obviously a highly divisive figure.)
10) True, Trump’s approval rating in Michigan’s exit poll on Tues was only 44 percent, pretty much what the GOP’s gov candidate got. But that could be as much a case of the state GOP dragging him down as vice versa.
11) Also, keep in mind that Trump only had a 39 percent favorable rating in Michigan when he won in 2016.
12) With a little course correction—more on trade, manufacturing, infrastructure; laying off tax cuts for the wealthy; not trying to undermine health insurance protections—I’d guess Trump could get back to his 2016 margin among working-class whites.
13) If he does that, Michigan is basically a tossup. I re-calculated Tuesday night’s vote, except with Trump’s 2016 white working-class margin rather than the 2018 margin, and it produces a 51-49 win for Dems (actually a bit narrower).
14) It’s probably obvious why: the white working-class is 40-50 percent of the state’s electorate. White women with college degrees are only about 15 percent.
15) Now Democrats may also have room to grow in 2020—they could boost turnout among African American voters, whose turnout on Tuesday was roughly equivalent to their 2016 turnout (suggesting a combination of a strong midterm turnout and a weak 2016 turnout).
16) But the bottom line is that Michigan is very likely to be a slog in 2020. Democrats are going to have to pour a lot of resources into it and other Midwestern states to lock them down. There is no wall.
17) An addendum: As people have pointed out, the ballot initiative that passed in Mich on Tuesday--including same-day voter registration on some related measures--should help Dems. And a recession could obviously turn a close race into a blowout.
